PLANES PARA LA TEMPORADA DE 2007

Human behavior is argued to be one of the lead contributors to our modern ecological problems. Thus, viewing behavior change as a solution for resolving environmental issues is logical (Beretti et al., 2013). It should be noted that behavioral solutions for our environmental problems should be explored in conjunction with other

56

possible tools, such as regulatory and technological innovations. Behavioral change can be related to the term ‘behavioral capital’, which refers to the notion that altering

citizens’ behaviors can affect improvement in environmental quality (Beretti et al., 2013). Behavioral research has found that behaviors can be explained and interpreted through underlying norms and interests that often compete with one another. The dual interest

metaeconomics framework includes a theory that recognizes two human tendencies that

act as motivating forces in behavior: egoistic-hedonistic based self-interest and empathy- sympathy based other-interest (Czap et al., 2012). The egoistic-hedonistic self-interest contributes to behaviors that enhance financial and personal utility; relating to a private dimension. The empathy-sympathy other-interest is a force that contributes to

“stewardly” or “social” behaviors, which improve the community. The adoption of pro- environmental behaviors puts an emphasis on the importance of the other-interest because mitigating environmental problems through individual behavior will benefit the public, rather than just the self. These internal forces coexist in the individual and help determine intentions in decision-making.

The dual-interest theory is relevant to the adoption of community solar because it forces individual decision-makers to evaluate the tradeoffs of community solar

enrollment. Participating in community solar includes both self and other-interest benefits, but they are temporally-dependent. Self-interest puts value on how much a good, service, or behavior increases the utility of an individual. This utility relates to the perceived costs and benefits associated with adopting the particular product or behavior. Financial motives are most commonly considered self-interest forces. The financial

57

dimension of participating in community solar may be perceived as damaging to the economic self-interest, because the up-front cost of enrollment is more expensive than simply opting to use the default utility electricity. However, the economic value of community solar enrollment may be heightened when the time frame is considered; an individual may save money in the long term if the bill credit rate is substantial enough (the bill credit is the rate at which community solar members are paid back for their proportional share of solar electricity generation). More specifically, the return on investing in a community solar project may have a relatively short pay-pack period, which will ultimately benefit the economic self-interest of participating individuals. Other personal benefits of community solar participation involve the Renewable Energy Certificates owned by the participants as well as potential tax advantages. Therefore, self- interests such as long-term financial incentives can be significant motivating factors in community solar enrollment.

Participating in renewable energy projects also benefits the community and the environment. Thus, the other-interest may act as a strong motivating factor in community solar participation. Enrolling in a community solar project will help bring more

renewables online for the utility, which will consequently supplant coal-fired electricity in the grid. Reducing fossil-fuel generated energy subsequently decreases the emissions (carbon, methane, and pollutants) associated with generation, and therefore aids in climate change mitigation. Climate change globally impacts all communities, thus engaging in a project or behavior that potentially lessens the effects of such a harmful phenomenon can be considered a socially and environmentally beneficial action.

58

Residents who perceive climate change as a serious threat and who hold a positive attitude about renewable energy may also contain a strong other-interest motivation to join a community solar project. As both internal motivators will likely come into play when residents are making a decision about enrolling in community solar, using the dual- interest framework will be helpful in conceptualizing how self-interests and other-

interests shape attitudes towards renewable energy and potential interest in enrolling.

3.2 METHODS

3.2.1 STUDY AREA

The City of Portland has the largest population of residents eligible for participation in community solar projects in the state of Oregon, with a population of 632,309 residents (U.S. Census, 2015). Portlanders are served by two Investor-Owned Utilities (IOUs): Portland General Electric (PGE) and Pacific Power, the electricity providers that will be developing community solar programs for their customers. PGE is the largest IOU in the state of Oregon, with most of its service territory located in Portland. Pacific Power serves a smaller proportion of Portland, but has pockets of other territories throughout Oregon and five other Western states. The climate of Portland is mild and cloudy for portions of the year, where sunshine is rare during most of the year. The usage of solar power in the form of community solar in Portland may seem infeasible to many residents due to perceived lack of sun resources. Assessing interest in

59

resource in this region, is important to the development of a successful community solar program.

Another reason I deemed Portland as an appropriate study area for this research is because the city has been a pioneer in climate action planning. The Portland Climate Action Plan (CAP) is a comprehensive document that provides mitigative and adaptive strategies for residents, industry, and businesses to adopt to address climate change. These strategies are suggested to help Portland accomplish an 80 percent reduction in local carbon emissions by 2040 (Anderson et al., 2015). There’s a brief section in the Portland CAP that suggests that community solar could be one mechanism to help

achieve this goal. Therefore, assessing the familiarity of Portland residents in terms of the Portland CAP will possibly reveal the baseline awareness of this plan and whether

residents are aware that community solar is considered a climate change mitigation tool.

3.2.2 SURVEY DESIGN

The primary qualitative and quantitative data used in this research were collected through a questionnaire designed for Portland residents. Much of the information used to develop the survey questions and to provide context for Portland’s energy landscape was collected through key informant interviews. These interviews were semi-structured in format and captured information about the legislative and regulatory context for community solar in Oregon, the program details such as location, cost, and size, the potential barriers hindering adoption in Portland, and general information on the utility structure and system in Portland. Key informants were selected by researching active

60

organizations and participants involved in the community solar rulemaking process as well as local solar companies and environmental nonprofits. Six key informant interviews were conducted and recorded, each lasting roughly an hour.

The questionnaire, organized into six sections, contained a total of 38 questions, which were divided into categories on general energy, solar energy, community solar, climate change, energy use, and demographics. Most questions used a 5-point Likert- scale, ranking, multiple choice, or text-entry structure. The questions in the survey that measured variables relevant to this study inquired about knowledge, awareness, perceived barriers, energy attitudes, climate change beliefs, community involvement, and peer influence. The dependent variable was measured through a question asking the respondent to rate their level of interest in joining a community solar project.

Awareness, familiarity, and knowledge of energy topics were evaluated in six questions. Awareness was measured using a 5-point Likert scale that asked respondents to report their level of familiarity with four solar options (from ‘Not at all familiar’ to ‘Very familiar’): voluntary green utility programs, community solar programs, leasing rooftop solar panels, and owning rooftop solar panels. Further measurement of

community solar awareness was conducted through a question asking respondents to report how much prior knowledge they had about community solar (from ‘Never heard of them’ to ‘I know a great deal’). Knowledge was assessed in three questions, asking respondents to rate their knowledge level in terms of renewable energy in general and solar energy. Further, I asked respondents to choose the correct proportion of renewable energy (non-hydro) in Portland’s electricity mix to gauge how knowledgeable they were

61

about Portland’s usage of renewable energy. Finally, I asked my survey respondents to rate their awareness level of the Portland CAP.

Attitudes and beliefs towards energy and climate change were other variables I wanted to test against my variable of interest. I provided positive and negative statements about renewable energy, climate change, and local impacts of climate change and then asked respondents to rate their level of agreement. Further, to assess whether there are significant perceived barriers to utilizing solar energy among the Portland market, I included a bank of barrier statements and asked respondents to rate their agreement. The purpose of capturing the barriers of solar energy was to investigate what may possibly hinder the diffusion of community solar and other forms of solar energy in Portland, whether it’s a lack of information available or for economic reasons. Lastly, elements of social influence and community engagement were measured to examine whether interest in community solar was heightened among residents actively involved in community activities or if they knew their peers were participating in community solar as well.

3.2.3 SAMPLING & IMPLEMENTATION STRATEGY

As I wanted to include a gradient of demographics within my survey sample and I wanted to cover all geographic regions of Portland, I chose to stratify my sample by choosing two neighborhoods within each of Portland’s seven neighborhood districts. These neighborhood districts each have their own specific neighborhood coalition: Central Northeast Neighbors, East Portland Neighborhoods, Northeast Coalition of Neighborhoods, North Portland Neighborhoods, Neighbors West-Northwest, Southwest

62

Neighbors Inc., and Southeast Uplift Neighborhoods. Fourteen Portland neighborhoods were included in the study area to capture a gradient of demographics. Some

neighborhoods were lower-income, while others had higher median incomes. As

community solar is an energy practice intended to be inclusive of all economic classes, I wanted to include a variety of neighborhoods that had varying economic profiles. I selected the 14 neighborhoods in my sample by numbering every neighborhood listed under each district and then randomly selecting two neighborhoods using a random number generator. The neighborhoods selected were Linnton, Hillside, Kenton,

Overlook, Sabin, Sullivan’s Gulch, Rose City Park, Madison South, Russell, Hazelwood, North Tabor, Foster-Powell, Multnomah, and West Portland Park.

To select eligible households randomly in each neighborhood, I used imagery from Google Maps to number every residential street in each selected neighborhood. Using a random number generator, I randomly selected five streets in each neighborhood and then numbered each home on every selected street and randomly selected 15

households per street. I recorded the addresses of each selected home in a household database. Thus, there were 75 homes per neighborhood, totaling 1,050 households. This sampling strategy was used to comply with Dillman’s (2000) suggestions for

representative sample sizes. Eligible survey participants were Portland residents over the age of 18 who were the primary utility-bill payers of their household. These residents were either customers of PGE or Pacific Power, as well as homeowners or renters.

The questionnaire was distributed from September to December 2016 using a hybrid distribution approach. Surveys were delivered to households via a modified

63

“Drop-Off/Pick-Up” method (Steele et al., 2016) for the first two rounds of contact. In teams, survey packets containing the paper questionnaire, cover letter, and business reply envelope were either hand delivered to survey participants at their home or dropped off on their front door (if participant was not home) during the first round of contact. One to two weeks later, if a participant had not completed a survey and if they hadn’t refused participation, a door-hanger was given to them as a reminder. The survey delivery methodology then transitioned to the Tailored Design Method (Dillman, 2000) for the last two rounds of contact, mailing survey packets and postcards as additional reminders. The survey modes offered were paper or web-based. Surveys were retrieved via three different approaches: picked up in person from the participants’ home two days after drop-off (completed survey was left on the front door in a plastic bag provided by the research team), mailed back in a business reply envelope, or completed online.

3.2.4 DATA ANALYSIS

Exploratory Factor Analysis (EFA) was the process I used to explore the structures of various attitudes and beliefs in my survey in order to create aggregated response variables for logistic regression analysis. EFA is an appropriate tool to use for survey data analysis because it aids in understanding the underlying dimensions of bank- structured survey questions that contain multiple statements. It’s important to be aware that EFA comes with several assumptions, such as a large sample size, tested using the Kaiser-Meyer-Olkin index, correlated data (checked through the Bartlett test of

sphericity), and multivariate normality. This last assumption cannot be satisfied through Likert-style survey data, thus there are risks associated with the instability of parameter

64

estimates. However, given the usefulness of EFA in extracting factors for use in regression, I deemed it a suitable mechanism for analysis.

I utilized EFA for three different questions in my survey to explore whether variables loaded on to each other to form factors. Specifically, I conducted EFA to investigate attitudes towards energy, hypothesizing that my statements would capture “positive”, “negative”, “indifferent” and “expensive” beliefs about energy. Additionally, I used EFA to explore attitudinal dimensions of climate change beliefs, as well as

perceived barriers of solar energy use. To select factors among these variables, I used both a scree test and parallel analysis, and the “varimax” rotation for estimating the factors. Eigenvalues, which measure the variance in all the variables under the selected factor, were also calculated to assess the explanatory power of the factors. If variables loaded well on to each factor, then I created a new index of the mean responses for each question accounted for in the factor. These indices were then applied as independent variables in my attitude logistic regression model.

I constructed a logistic regression model to assess how interest in community solar was influenced by energy attitudes, awareness, knowledge, climate change beliefs, perceived barriers, and general interest in solar power. I used logistic regression analysis because I was able to transform my dependent variable into a binary response variable:

very interested or somewhat interested responses were coded as “Interested” (1) and all

other responses coded as “Uninterested” (0). Additionally, this analysis doesn’t contain stringent rules regarding normality or data continuity. Following the Aikake Information Criterion (AIC), I built a reduced attitude model by eliminating variables from the full

65

model that increased the model AIC (a backward step-wise approach). To check for multi-collinearity issues among variables in the model, I calculated the Variance Inflation Factor (VIF). Pseudo r-squared (Nagelkerke r-squared) was also calculated to assess the predictive power of each model. I converted the slope coefficients of each variable into odds ratios to assess the strength of each variable on predicting interest in community solar. Descriptive analyses were also conducted to report the response frequencies, central tendencies, and spreads of the variables used in the attitude models. All statistical analyses were performed in R Studio (Version 3.3.2).

3.3 RESULTS

3.3.1 RESPONSE SUMMARY

I received a total of 330 completed questionnaires yielding a response rate of 34.2%. Table 3.1 displays the scale, central tendency, spread, and count of each item evaluated in the attitude logistic models.

Most survey respondents reported being slightly or moderately knowledgeable about renewable energy in general, solar energy, and the proportion of renewable energy in Portland’s electricity mix. Knowledge about the Portland Climate Action Plan was much lower, in fact almost 85% of respondents reported being not at all or not very informed about the Portland CAP. Energy conscientiousness was also measured, where

66

Table 3.1. Summary of items used in attitude logistic model. Energy attitudes, climate change beliefs, and solar barriers were truncated to form new response variables used in the model.

Factor Scale* Mean SD N

Knowledge

Renewable Energy 0-4 1.511 0.783 327

Solar Energy 0-4 1.505 0.785 329

PDX Renewable Energy Proportion 0-2 1.07 0.85 275

Climate Action Plan 0-4 0.72 0.78 319 Conservation

General Conscientiousness 0-4 2.340 0.634 329 Energy Attitudes

Aids in preventing CC -2 to 2 1.206 1.103 326

Helps transition away from fossil fuels -2 to 2 1.332 0.981 325

Reduces our impact -2 to 2 1.388 1.033 327

Continue use of fossil fuels because they’re

cheaper -2 to 2 -1.07 1.11 327 Makes no difference for CC -2 to 2 -1.083 1.152 324

Solar & wind are costly -2 to 2 0.196 1.019 326

Plenty of fossil fuels left -2 to 2 -0.541 1.272 327

Renewable energy is overrated -2 to 2 -1.492 0.930 323

Maintenance & installation are costly -2 to 2 0.463 0.946 326

Too busy to think about it -2 to 2 -1.022 1.017 325

Never comes to mind -2 to 2 -0.884 0.993 327

Don’t care as long as it’s affordable -2 to 2 -1.206 0.977 325 Awareness

Voluntary green utility programs 0-4 1.673 1.261 327

Community Solar 0-4 0.688 0.710 327

Leasing rooftop solar panels 0-4 0.789 0.976 327

Owning rooftop solar panels 0-4 1.220 1.077 327

Prior knowledge about community solar 0-4 0.633 0.757 324 Climate Change Beliefs

Not as bad as it’s portrayed -2 to 2 -1.188 1.172 314

Nothing we can do to stop it -2 to 2 -0.981 1.173 315

It’s a natural phenomenon -2 to 2 -0.578 1.331 313

It’s a hoax and conspiracy -2 to 2 -1.786 0.686 313

Dire consequences for all life -2 to 2 1.637 0.794 317

It’s caused primarily by humans -2 to 2 1.194 1.095 314 Barriers to solar use

Roof not suited -2 to 2 -0.06 1.24 327

No time -2 to 2 -0.49 1.17 317

Rent my home -2 to 2 -1.17 1.54 305

Not interested -2 to 2 -1.14 1.05 322

Costs too high -2 to 2 0.33 1.09 321

Too much hassle -2 to 2 -0.27 1.12 321

Lack of knowledge -2 to 2 0.26 1.21 321

Planning on moving soon -2 to 2 -0.23 1.38 323

Too new to the market

Concerns about reliability & maintenance

-2 to 2 -2 to 2 -0.74 0.24 1.02 1.24 323 325 Community Peer Influence 0-4 1.90 1.12 319

67

Community involvement**

Interest- dependent variables 0-7 1.88 1.99 329

Interest in solar energy 0-4 2.56 0.96 327

Interest in community solar

Timing of community solar adoption

0-4 0-4 2.66 2.53 0.99 0.89 324 321 *Scale definitions for each variable can be found in Appendix A. **Community involvement was calculated by adding number of days spent each year on a community activity and then coding (0 to 7) based on frequency.

over 90% of respondents rated themselves as somewhat or very energy conscious. Awareness levels were low among respondents: almost 90% of respondents were unfamiliar with community solar, 81% were unfamiliar with leasing solar panels, 62% were unfamiliar with owning solar panels, and 47% were unfamiliar with voluntary green utility programs. 52% of respondents had no prior knowledge about community solar before participating in the survey (34% had a little).

For attitudes towards renewable energy, the majority of respondents agreed with positive statements about energy and disagreed with negative energy statements. Neutral responses were common for statements reflecting the belief that renewable energy is too expensive to utilize. Statements that stressed indifference towards renewable energy were also unpopular among survey respondents. Beliefs towards climate change followed a similar environmentally-leaning pattern; most respondents disagreed with statements that paralleled a denial attitude, while agreed with statements that stressed the seriousness of climate change.

Perceived barriers of solar energy use were not extreme: most respondents reported neutral or negative responses to barrier statements. Though the two barriers that had the highest proportion of respondents agreeing with them were economic and lack of knowledge barriers. Most respondents disagreed with the barrier stating they were not

68

interested in pursuing home solar, which is in line with the finding that 50% of

respondents reported being very or extremely interested in having access to solar energy.

Most respondents felt that if they knew a peer who was enrolled in a community solar project, they would be somewhat more likely to join. About 27% of respondents believed their likelihood of joining would be a lot higher if they knew someone already enrolled, while ~33% believed their enrollment decision wouldn’t be influenced if they